Hi Claudia,

Claudia De Bartolomeis wrote:
Hi everyboby,
I'm designing a custom board based on M5329EVB and the BSP m532xevb based on kernel uClinux 2.6.17.7-uc1 by Freescale. M5329EVB Sdram is 32MB.
My custom board has a 64MB SDRAM.
To configure the ram size I changed the built time option
CFG_SDRAM_SIZE from 0x2000000 to 0x400000.
I modified also chip select in the file /arch/m68knommu/platform/532x/config.c, function sdram_init()
adding the line :
/* Initialize SDRAM chip select */
  MCF_SDRAMC_SDCS0 = (0
   | MCF_SDRAMC_SDCS_BA(SDRAM_ADDRESS)
   | MCF_SDRAMC_SDCS_CSSZ(MCF_SDRAMC_SDCS_CSSZ_64MBYTE));
It works but accessing the SDRAM becomes 8 time slower!!!

How did you measure that it was 8 times slower?

In any case I would suspect a cache setup problem.


I have no idea what could have caused this side effect,
but I have to notice, that in original M5329EVB board bsp, the function sdram_init(), would have done nothing,
because SDRAM initialization is perfomed by the bootloader dBUG.
Is it necessary to inizialize the SDRAM in the bootloader, or is it also

Yes - if the kernel is loaded into RAM for execution.
If your kernel runs from flash then it can/would do it.


possible do that into the kernel (sdram_init())?
Could the slow access to SDRAM  be due to that?
Or does it depend on something else it is missing or wrong?

Check the cache setup code. It may not be set to cache over
the new larger range you have. Check the code in
linux-2.6.x/include/asm-m68knommu/mcfcache.h

Regards
Greg


------------------------------------------------------------------------
Greg Ungerer  --  Chief Software Dude       EMAIL:     [EMAIL PROTECTED]
SnapGear -- a Secure Computing Company      PHONE:       +61 7 3435 2888
825 Stanley St,                             FAX:         +61 7 3891 3630
Woolloongabba, QLD, 4102, Australia         WEB: http://www.SnapGear.com
_______________________________________________
uClinux-dev mailing list
uClinux-dev@uclinux.org
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by uclinux-dev@uclinux.org
To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev

Reply via email to